AFDBench is a new benchmark that evaluates how well large language models can generate professional Area Forecast Discussions (AFDs) for the National Weather Service by reasoning through structured AI weather forecast data. It contains 7,732 expert-written discussions paired with real forecast inputs and introduces three metrics—Met-Align, Style-Align, and Input-Grounding—to assess numerical accuracy, professional dialect adherence, and fidelity to source data. Zero-shot tests show open-source LLMs perform poorly on style and grounding, but reinforcement learning with Group Relative Policy Optimization nearly doubles style alignment and improves grounding, enabling a 7B-parameter model to write like a professional meteorologist.
arXiv:2608. 09948v1 Announce Type: cross Abstract: No single AI weather model excels at all variables, pressure levels, and lead times.
arXiv:2607. 23983v1 Announce Type: cross Abstract: Operational flood forecasting depends on tacit forecaster expertise that is difficult to formalize, audit, and transfer.
The paper presents Laxmi, a retrained version of the AIFS weather model that uses satellite-based precipitation observations instead of ERA5 reanalysis data. Laxmi achieves a 19% improvement in global probabilistic accuracy, reduces drizzle overprediction by 33%, and boosts the 95th percentile Brier skill score by 57%. In a case study of 10 Indian tropical storms, Laxmi accurately forecasted 150 mm event-total precipitation in 7 events, outperforming both the original AIFS and the leading physical model IFS.
arXiv:2606. 26421v1 Announce Type: new Abstract: State-of-the-art medium-range AI weather models can outperform traditional Numerical Weather Prediction (NWP) but require massive training budgets.
The paper introduces SmartWeatherAgent, a three‑stage architecture that combines intent recognition, hazard prediction, and reasoning‑enhanced generation to improve tourism meteorological services. It fuses rule‑based methods with large language models and a LightGBM model enriched with highland‑specific features, achieving an F1‑Macro score of 0.605 and 1.60 ms latency on high‑wind, precipitation, and low‑temperature events. A 12‑round micro‑step prompt self‑optimization loop raises the composite warning quality score from 4.2 to 8.9, with notable gains in data source citation, physical mechanism explanation, and scientific rigor through explicit uncertainty statements.
The article discusses how AI foundation models trained on weather and climate data are increasingly fine‑tuned for a wide range of Earth science tasks beyond traditional weather forecasting. It highlights that current evaluations rely mainly on benchmark skill metrics, which assess forecast accuracy but not the models’ representation of underlying physical processes, a gap that becomes critical under changing climate conditions. The authors propose five priorities for physical evaluation—covering training data, fine‑tuning, behavioral testing, mechanistic interpretability, and output validation—and recommend three actions for the next decade: open AI‑ready evaluation datasets, a shared physics‑based reporting standard, and a dedicated research program on model safety.
arXiv:2606. 02663v1 Announce Type: cross Abstract: Recent advances in machine learning have produced probabilistic weather forecasting models comparable to state-of-the-art numerical weather predictors.
The paper introduces Probabilistic Bias Correction (PBC), a machine learning framework that learns to correct historical probabilistic forecasts, thereby reducing systematic errors in subseasonal weather predictions. Applied to leading dynamical and AI models from ECMWF, PBC doubles the AI system’s modest subseasonal skill and improves the operationally-debiased dynamical model for most pressure, temperature, and precipitation targets. In ECMWF’s 2025 real‑time forecasting competition, PBC’s global forecasts ranked first across all weather variables and lead times, outperforming multiple operational and ensemble models.
CastFlow is a role‑specialized agentic forecasting framework that combines general‑purpose tool reasoning with specialized temporal forecasting. It uses an Experience Memory to guide a frozen LLM in acquiring evidence via a multi‑view toolkit, then a lightweight specialized LLM refines predictions through evidence‑grounded reasoning. The two‑stage training strategy—supervised fine‑tuning followed by reinforcement learning with verifiable rewards—optimizes the specialized model based on final forecasting outcomes, achieving superior performance on nine out of ten real‑world benchmarks.
arXiv:2604. 16238v2 Announce Type: replace Abstract: Decision-makers rely on weather forecasts to plant crops, manage wildfires, allocate water and energy, and prepare for weather extremes.
State-of-the-art medium-range AI weather models can outperform traditional Numerical Weather Prediction (NWP) but require massive training budgets. This restricts usage for under-resourced groups and severely limits fast model iteration.
arXiv:2604. 12306v3 Announce Type: replace-cross Abstract: Climate decision-making in the GCC states increasingly demands systems that can translate heterogeneous scientific and policy evidence into actionable guidance, yet general-purpose large language models (LLMs) remain weak both in region-specific climate knowledge and grounded interaction with geospatial and forecasting tools.